[nos-bbs] netrom related crashes (INP)

[maiko at pcsinternet.ca]

> Can you tell me a bit about HOW the ‘crashes’ you’re talking about
> present themselves?

GDB just breaks out and dumps stuff, that if you look at the source
code names and line entries showing up, most of them land in INP2011
defined areas of the source. For instance :

#0  0x00007f30f855a780 in __memcmp_avx2_movbe () from /lib64/libc.so.6
#1  0x0000000000494a90 in addreq (
     a=0x55aa55aa55aa55c1 <error: Cannot access memory at address 
0x55aa55aa55aa55c1>, b=0x4f3587 <Nocall> "\234\236\206\202\230\230`") at 
ax25subr.c:228
#2  0x00000000004b0431 in callnocall (
     nodecall=0x55aa55aa55aa55c1 <error: Cannot access memory at address 
0x55aa55aa55aa55c1>) at newinp.c:235
#3  0x0000000000481560 in nr_bcnodes (ifp=0x55aa55aa55aa55aa) at 
nr3.c:673
Backtrace stopped: previous frame inner to this frame (corrupt stack?)

OR

#0  0x00007f8899b47780 in __memcmp_avx2_movbe () from /lib64/libc.so.6
#1  0x0000000000482ab5 in nrresetlinks (rp=0x110d720) at nr3.c:1432
#2  0x000000000047cbf0 in doobsotick () at nrcmd.c:1293
#3  0x0000000000451b34 in timerproc (i=0, v1=0x0, v2=0x0) at timer.c:59
#4  0x0000000000455bc4 in _kicker (func=0x451a7f <timerproc>, iarg=0,
     parg1=0x0, parg2=0x0) at ksubr.c:143

Recompiling clean (#undef INP2011 in config.h) has stabilized it 
significantly.

The INP2011 code was written long ago, and I just don't have the time
to spend on it right now, so quite far down on the priority list :(

> On my pi, if one watches my jnos F9 screen and runs ’top’ along with
> ‘jnos', I see, in almost every single crash, CPU usage goes to 100%
> for jnos on the ’top’ screen and jnos locks up.  Is this true of the
> system crashes you are experiencing? Killing jnos from another
> terminal is the only way out.

Nope, mine just outright crashes, no lockup.

Maiko / VE4KLM